Impact of a dividend initiation wave on shareholder wealth

Abstract

This article examines the short- and long-run impacts of a dividend initiation (DI) wave period on shareholders’ wealth. I test two hypotheses. First, firms initiating dividend payments during a DI wave period experience lower announcement returns than those initiating dividend payments outside a DI wave period. Second, firms initiating dividend payments inside a DI wave period underperform those initiating dividend payments outside a DI wave period in the long run. Using a sample of 688 DI announcements from the period 1977 to 2010, I find evidence supporting both hypotheses. Since a firm’s decision to initiate a dividend payment during a DI wave period can have implications on its shareholders’ wealth in both short and long runs, the results of this study can help design investment strategies.

Keywords:

• dividends
• dividend initiation waves
• dividend signalling
• CARs
• BHARs

JEL Classification:

• G30
• G35

Acknowledgements

The author would like to thank two anonymous reviewers, his adviser, Dr Chintal Desai, and other faculty in The University of Texas – Pan American (UTPA) Economics and Finance Department for their valuable comments and suggestions. The author is solely responsible for errors and omissions in this article.

Notes

¹ For example, Asquith et al. (1983) show that the stock market reacts positively to a firm’s DI announcement. Pettit (1972), Aharony and Sway (1980), Denis et al. (1994) and Elfakhani (1998) document that the stock market reacts positively (negatively) to dividend increases (dividend decreases). Healy and Palepu (1988) show a negative reaction to the dividend omission announcements.

² This method was first used by Mitchell and Mulherin (1996) and then later on formally developed by Harford (2005).

³ For a comprehensive review on dividends, see Allen and Michaely (2003) and DeAngelo et al. (2007).

⁴ The authors used 2-day window (–1, 0) surrounding announcement date.

⁵ According to Welch (1992), bandwagon effects may appear if potential investors pay attention not only to their own information about an IPO, but also to whether other investors are purchasing.

⁶ Loughran and Ritter (1995, p. 24) examined the long-run IPOs performance and stated that ‘… the degree to which issuing firms underperform varies over time: firms issuing during years when there is little issuing activity do not underperform much at all, whereas firms selling stock during high-volume periods severely underperform …’. My long-run performance hypothesis (H2) is similar to this line of thinking.

⁷ The regular cash dividend has a CRSP distribution code from 1200 to 1292.

⁸ In general, any DI is announced within 1 year after IPO date will be announced in IPO prospectus.

⁹ Table 1 is reproduced from Desai et al. (2013) with the authors’ permission.

¹⁰ I categorize the sample into 10 industries based on SIC codes provided by Professor Kenneth R. French’s website. Professor Kenneth R. French originally separate into 12 industries. Because of the remove of Financial and Utilities industries, I have now 10 industries in the study sample.

¹¹ Utilities and finance industries are not reported because there is no firm in the sample belonging to these industries.

¹² For example, Biomet Inc. was not found in Datastream, but was found in Mergent Online database. Return-on-asset ratio and book value of equity in market-to-book ratio data are collected from there.

¹³ Data are collected from the CRSP database.

¹⁴ Data for ROA and MTB are collected from Datastream and Mergent Online database.

¹⁵ Calculated by taking SD of daily market returns 1 year prior to its DI announcement. Daily market returns are collected from Professor Kenneth R. French’s website.

¹⁶ Data for Turnover, Mature, StockVolatility and MarketReturn are collected from the CRSP database.

¹⁷ I follow Desai et al. (2013) to define high-tech firms.

¹⁸ I exclude data for 6 days prior to an announcement date to prevent possible information leakage.

¹⁹ The data to run this model are collected from CRSP and Professor French’s website.

²⁰ I also test with equally weighted index and the results are consistent and available upon request.

²¹ I also perform cross-sectional regression using robust SEs clustered at the industry level (Nichols, 2007). The results are similar and available upon request.

²² Bulan et al. (2007) use the logarithm of assets of the firm 1 year prior to its DI announcement as a proxy for the size.